Speaker Interview: Thomas Munro

Parallelism in PostgreSQL 11   Friday 11:30   Hotel

Twitter: @MengTangmu Blog: write-skew.blogspot.com Company website: enterprisedb.com

Could you briefly introduce yourself?

I am Thomas Munro, a New Zealand-based software developer working for EnterpriseDB. I’ve been working on PostgreSQL and some related things for about 4 years now and recently became a committer. Before that I did a decade or two of proprietary application development work, mostly involving databases and networks and Unix and C++, mostly in Europe.

How do you engage with the PostgreSQL Community?

Primarily by exchanging code, reviews and discussion on the pgsql-hackers and pgsql-bugs mailing list. I also try to get to a couple of the big overseas conferences each year and meet developers and users, and I am very lucky that my employer supports me in that. Finally, there is a new Australia/New Zealand association PGDU.org that I’m involved in; we’re trying to grow the active community with an annual mini-conference in this far corner of the world. Come on down next year, it’s in our summer and we have good beaches.

Have you enjoyed previous pgconf.eu or FOSDEM conferences, either as attendee or as speaker?

I went to three FOSDEMs as an attendee when I lived nearby. I really enjoyed it, and it was an important part of my path to working on open source software; not only in an abstract eye-opening sort of way, but also because I met a future colleague in the beer queue there! I haven’t made it to pgconf.eu yet (fingers crossed).

What will your talk be about, exactly? Why this topic?

PostgreSQL’s new ability to use more than one CPU core for a single query. I worked on a couple of features in this area, building on work from many other people, and I know from personal experience that the new concepts and terminology are confusing and mysterious when you first encounter them. My goal is to give the audience an intuitive understanding of what’s going on, how and when parallelism can help you, and also some motivation and context for this work.

What is the audience for your talk?

Pretty much any user of PostgreSQL, experienced or not, since this is all new magic and some may even have misconceptions from other relational databases.

What existing knowledge should the attendee have?

Intermediate level SQL and familiarity with EXPLAIN.

What is the one feature in PostgreSQL 11 which you like most?

It’s hard to choose one, but I think JIT-compiled expressions are exciting, not so much for the results today, but because they are the tip of another iceberg.

Which feature would you like to see in PostgreSQL?

I could give so many answers... Here’s something my colleagues and I are working on: a new MVCC implementation using in-place-updates and undo logs, instead of vacuum.

FOSDEM is a very large conference. Are there any other talks you want to see? Where will people usually find you?

I’ll also be talking about PostgreSQL-on-FreeBSD in the BSD dev room. FOSDEM is so big and the schedule isn’t up yet so it’s very hard to say where else I might lurk (always difficult choices!), but if it has a database or a kernel or a compiler in it I’m definitely interested. Also, the hallway track is so important. Looking forward to meeting you there!